EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Multi-Objective Hyperparameter Optimization of Lithium-Ion Battery State-of-Health Prediction Models

Andrés Barrantes Boza, Juan José Montero Jiménez, Rob Vingerhoeds

Applied Sciences · 15.11.2026 · 16 (22) · 11321

DOI: 10.3390/app162211321 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 15.11.2026. Verfasst von Andrés Barrantes Boza und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Selecting machine learning models for state-of-health (SOH) prediction in predictive maintenance involves trade-offs among competing criteria, such as predictive accuracy, robustness to measurement noise, and computational requirements. Although model comparisons are commonly based primarily on prediction error, this approach may overlook characteristics relevant to practical deployment, particularly in resource-constrained systems. This work proposes a multi-objective hyperparameter optimization framework for SOH prediction based on the nondominated sorting genetic algorithm II (NSGA-II), in which three commonly used prediction models are evaluated and compared: support vector regression (SVR), long short-term memory (LSTM), and physics-informed neural networks (PINNs). For each model, NSGA-II optimizes model-specific hyperparameters with respect to validation RMSE, sensitivity to measurement noise, and inference time. Noise sensitivity is quantified as the relative increase in RMSE across six Gaussian-noise levels ranging from 5% to 30%. The framework is evaluated through a case study using the lithium-ion battery aging dataset from the NASA Prognostics Center of Excellence. The resulting Pareto-optimal solutions quantify the trade-offs involved in model selection and support practitioners in identifying configurations that match their specific requirements.

Zitieren

APA

Boza, A. B., Jiménez, J. J. M., & Vingerhoeds, R. (2026). Multi-Objective Hyperparameter Optimization of Lithium-Ion Battery State-of-Health Prediction Models. Applied Sciences, 16(22), 11321. https://doi.org/10.3390/app162211321

Vancouver

Boza AB, Jiménez JJM, Vingerhoeds R. Multi-Objective Hyperparameter Optimization of Lithium-Ion Battery State-of-Health Prediction Models. Applied Sciences. 2026;16(22):11321. doi:10.3390/app162211321

BibTeX
@article{Boza2026,
  author = {Boza, Andrés Barrantes and Jiménez, Juan José Montero and Vingerhoeds, Rob},
  title = {Multi-Objective Hyperparameter Optimization of Lithium-Ion Battery State-of-Health Prediction Models},
  journal = {Applied Sciences},
  year = {2026},
  volume = {16},
  number = {22},
  pages = {11321},
  doi = {10.3390/app162211321}
}
Kostenlos starten →